GOOD HOUSEKEEPING
Make choosing a meal
part of the experience.
I led the product vision for Good Housekeeping’s recipe app, helping turn a large catalogue into an experience built around choosing and cooking a meal. With the team, I used prototypes to explore swipe-based discovery and give senior stakeholders something concrete to shape together.
The recipe data set limits on what we could build, so I helped align the experience with what the catalogue could reliably support. Later, I built a proof of concept for AI-led cleanup and enrichment, and proposed a small data pilot before committing to new features.
The constraints
The experience had to work
with the catalogue we had.
The discovery ideas depended on consistent recipe information. The historic catalogue had uneven metadata: some recipes had useful tags or nutrition information; others had generic categories or no usable tags. Cooking times could sit inside ingredient fields, with conflicting information elsewhere.
That made dependable filtering and other structured-data features difficult. Editorial capacity was limited, too. The team could not simply recreate or manually enrich the whole catalogue for the app.
I shaped the product around those realities, aligning stakeholders on what the available content could support rather than assuming every desirable feature was ready to build.
Returning to the problem
Test the data opportunity
before promising the features.
Later, I revisited the same constraint with a new capability: AI-led cleanup and enrichment. I built and iterated a proof of concept to make the opportunity tangible, then proposed a short data pilot before committing to new UX scope.
The order mattered. First establish what we could reliably extract and enrich from a small recipe sample; then decide which previously blocked features that could support. A cleanup pipeline would still need corresponding frontend work to become a useful product experience.
What changed
A defined product.
A demonstrated next possibility.
The wider team delivered an app combining swipe-based recipe discovery with step-by-step cooking. My contribution was the product direction, feature definition and prototypes that helped align the client around it.
The later enrichment proof of concept was built and shared. The larger cleanup and follow-on features remained a proposal, not a production rollout. It showed a plausible way back to ambitions the original data had put out of reach.